Introduction In cases of repeated treatment failure of periprosthetic joint infections (PJI) of the knee, above–the–knee amputation (AKA) or knee arthrodesis can be proposed to reduce the risk of recurrent infection, especially in cases with major bone defects or irreparable damage to the extensor mechanism of the knee. Since AKA versus knee arthrodesis results have been rarely assessed for these indications, we conducted a retrospective case-control study to compare both the rates of recurrent infection and functional outcomes. Hypothesis Patients who underwent AKA had fewer recurrent infections than those who had arthrodesis. Materials and methods Twenty patients who underwent AKA and 23 patients who had knee arthrodesis, between 2003 and 2019, were retrospectively included in this study. These two groups were comparable in age (73.8 versus 77.7 years (p = .31)) and sex (10 women and 10 men versus 16 women and seven men (p = .19)). Each group was analyzed individually and then compared in terms of survival (recurrent infection) and functional outcomes using clinical assessment scores (visual analog scale (VAS), French neuropathic pain questionnaire (DN4), Parker and Palmer mobility score and the 36-item short-form survey (SF–36)). Results The rate of recurrent infection was 10% (two out of 20 patients) for the AKA group and 21.75% (five out of 23 patients) for the arthrodesis group (p = .69). The mean follow-up for the AKA group was 4.18 years (1.2-11.8) and 9.7 years (1.1-14.33) for the arthrodesis group (p = .002). The number of previous revisions (three (1.5-4) for AKA and two (2-3) for arthrodesis) and the time between the primary arthroplasty and surgical procedure were significantly greater in the AKA group (48.0 (12.0-102.0) months) than the arthrodesis group (48.0 (24.0-87.0) months) (p < 001). The AKA group had significantly better clinical results for VAS (2.7 ± 2.2 vs. 3.1 ± 3.3), DN4 (1.5 ± 2.1 vs. 2.6 ± 2.9), Parker and Palmer (5.2 ± 1.7 vs. 4.6 ± 1.4), and SF–36 (30.9 ± 15.6 vs. 26.9 ± 17.0) (p < 001). Conclusion Above–the–knee amputation and knee arthrodesis showed no differences in the rate of recurrent sepsis. However, the comparison of the two groups demonstrated that patients who underwent an AKA had less pain, were more autonomous and had a better quality of life.